Understanding the Impact of the Low Interest Rate Environment on Retirement Security in the United States - A Review of Academic and Practitioner ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Aging and Retirement Understanding the Impact of the Low Interest Rate Environment on Retirement Security in the United States A Review of Academic and Practitioner Research May 2021
2
Understanding the Impact of the Low
Interest Rate Environment on Retirement
Security in the United States
A Review of Academic and Practitioner Research
AUTHOR Yimeng Yin * SPONSOR Retirement Section Research
Don Boyd * Committee
Hao Sun ** Retirement Section
*Center for Policy Research, Rockefeller College, University at Albany
**Department of Government and Public Affairs, Gallaudet University
Caveat and Disclaimer
The opinions expressed and conclusions reached by the authors are their own and do not represent any official position or opinion of the Society of
Actuaries or its members. The Society of Actuaries makes no representation or warranty to the accuracy of the information.
Copyright © 2021 by the Society of Actuaries. All rights reserved.3
CONTENTS
Executive summary ................................................................................................................................................... 4
Studies of interest rates: What caused the decline, and will it persist? ..................................................................... 4
Studies of how persistent low interest rates could affect the retirement security of individuals ............................ 5
Studies of how persistent low interest rates could affect pension funds, insurance companies, asset
management firms, and Social Security ....................................................................................................................... 6
1 Motivation and introduction .................................................................................................................................. 9
2 Studies of interest rates: What caused the decline, and will it persist? ................................................................. 10
2.1 The downward trend in interest rates since the 1980s....................................................................................... 10
2.2 What caused the decline in interest rates?.......................................................................................................... 10
2.3 Will the low interest rate environment persist? .................................................................................................. 13
2.4 Section conclusions ............................................................................................................................................... 17
3 Studies of how persistent low-interest rates could affect the retirement security of individuals .......................... 19
3.1 Sources of retirement income and how they are affected by low interest rates .............................................. 19
3.2 Important considerations in evaluating retirement security .............................................................................. 20
3.3 Studies of the impact of low interest rates on retirement security.................................................................... 22
3.4 Section conclusions ............................................................................................................................................... 28
4 Studies of how persistent low interest rates could affect pension funds, insurance companies, asset
management firms, and Social Security .................................................................................................................. 30
4.1 Defined benefit pension funds.............................................................................................................................. 31
4.2 Life insurance companies ...................................................................................................................................... 37
4.3 Defined contribution plans and asset management firms .................................................................................. 38
4.4 Social Security ........................................................................................................................................................ 39
4.5 Section conclusions ............................................................................................................................................... 40
5 Conclusions .......................................................................................................................................................... 43
Acknowledgments .................................................................................................................................................. 44
Notes ...................................................................................................................................................................... 45
Bibliography ........................................................................................................................................................... 47
About The Society of Actuaries ............................................................................................................................... 524
Executive summary
After decades of declines in real and nominal interest rates, intensified by the Great Recession, COVID-19, and
monetary policy responses, the United States has been and may continue to be in a protracted period of low
interest rates. Persistent low interest rates could affect the retirement security of Americans profoundly through
direct impacts on investment returns and through impacts on the capacities of governments and the private sector
to finance Social Security, pension funds, and other retirement savings mechanisms.
This report examines academic and practitioner research on these topics critically, with an eye toward drawing
common themes and conclusions, while reconciling and explaining discrepancies.
We have divided research into three broad topics:
• Studies of interest rates: What caused the decline, and will it persist?
• Studies of how persistent low interest rates could affect the retirement security of individuals, and
• Studies of how persistent low interest rates could affect pension funds, insurance companies, asset
management firms, and Social Security
STUDIES OF INTEREST RATES: WHAT CAUSED THE DECLINE, AND WILL IT PERSIST?
Real and nominal interest rates have been declining for decades. The declines were largely unanticipated by
economic forecasters and were a global phenomenon. This has important implications for retirement security. First,
if past retirement plans and saving decisions were based on expectations that interest rates and asset returns would
be higher than they are now, savers may not have accumulated enough wealth to support planned retirement
spending. Second, the worldwide decline in real interest rates makes it difficult, or impractically expensive, for
savers to offset the impact of the decline by holding a globally diversified portfolio of interest-bearing assets.
Several studies examined the historical decline from the related perspectives of how factors driving economic
growth affect interest rates, and how factors affecting desired savings and investment affect interest rates.
Slowing economic growth caused by slowing productivity growth and by slowing labor force growth as a result of
population aging has put downward pressure on interest rates by increasing the supply of loanable funds and
reducing the demand for investment. In addition, several factors directly affecting desired savings have placed
downward pressure on interest rates, including shifting demographics, rising inequality, and an emerging-market
savings glut. Factors directly affecting interest rates through desired investment, including a reduction in the price of
capital goods relative to labor and other inputs, lower public sector investment, and a rise in the spread between
risk-free rates and rates of return on capital from the early 1990s through the early 2010s also have put downward
pressure on rates.
Studies do not reach a consensus on how much of the interest rate decline is attributable to each factor but
generally agree that the above factors have been at work.
While there is still debate among economists, a consensus has been forming that the low interest rate environment
will be persistent. This can be seen in econometric forecasts from professional forecasters, and in projections
implied by financial markets. Despite the general agreement, specific projections of long-run equilibrium interest
rates vary across studies and are subject to great uncertainty.5 STUDIES OF HOW PERSISTENT LOW INTEREST RATES COULD AFFECT THE RETIREMENT SECURITY OF INDIVIDUALS The main sources of income and resources for U.S. residents age 65 or older are Social Security, income from earnings, pension income, income from financial assets, and real estate. The extent to which older individuals rely on these sources varies greatly by income range. Social Security is the largest income source by far for older individuals with below-median income and remains quite important for those between the median and the 75th percentile. Income from earnings is the most important source for individuals in the top quartile and is relatively important for the next quartile. Income from assets constitutes about 10 percent of total income for those above median income but plays a small role for those below the median. The potential impact of sustained low interest rates differs by type of income and therefore also differs by income group. Interest rates do not play a direct role in the calculation of Social Security benefits but could affect the ability or willingness of the federal government to pay full benefits, an issue we discuss in the next section. Income from work may not be directly affected by low interest rates but increased work is one possible behavioral response to lower interest rates. Payments from defined contribution plans and other investment accounts will be affected by how persistent low interest rates affect asset returns and risk. Several quantitative studies have examined the potential impact of persistent low interest rates. The estimated impact differs across studies, ranging from minimal to moderate depending on assumptions about asset accumulation and the treatment of home equity as a source of retirement income. One study estimated that low interest rates would have very little impact on income replacement rates upon retirement. However, this study assumed that people would (1) accumulate as much wealth in their working years, relative to their income, as previous generations had accumulated in higher interest-rate environments, presumably by saving more or by earning higher returns, and (2) borrow more through reverse mortgages when rates are lower, with no impact on overall financial well-being. (Munnell et al. 2013). By contrast, another study concluded that persistent low interest rates would have a larger negative impact on retirement security. (VanDerhei 2013). This latter study assumed that households would maintain their current saving pattern, which means a lower expected wealth-to-income ratio under the low-rate environment, and also that reverse mortgages are not available (or are not a free resource). The assumptions of VanDerhei (2013) lead to a more pessimistic outcome than those of Munnell et al. (2013). The studies we reviewed also found that the impact of persistent low interest rates differs across income and age groups. Although low-income households generally have worse retirement security compared to high-income households, they will face much less impact from the low interest rate environment because Social Security accounts for the majority of their retirement income and they have lower financial assets, and thus less risk of reduced investment income as a result of lower investment returns. (This assumes that Social Security payments are not at risk as a result of low interest rates.) Younger generations face a greater challenge in achieving retirement security in the low interest rate environment than older generations because they face a lower expected return on their retirement savings and longer expected retirement periods. Younger generations also face more favorable borrowing terms (e.g. lower rates for mortgages and auto loans) in low interest rate environments, but the studies we reviewed did not evaluate the extent to which lower borrowing costs can mitigate retirement security challenges (aside from the assumption of increased reverse mortgage borrowing in Munnell et al. (2013)). Studies also illustrated the important roles that behavioral responses, such as increasing savings rates and delaying retirement, can play in improving retirement security in a low return environment. The studies we reviewed suggest that most wage earners can achieve a reasonable replacement rate in a moderate low-return environment if they save between 10 and 15 percent of their income, start saving early, and save consistently throughout their careers (Blanchett, et al. (2018) and Byrne and Reilly (2018)). Although these savings rates are much higher than the savings rates required to achieve the same replacement rate under historical returns (usually in single digits), they can still
6 be considered in a reasonable range. Studies also have shown that delaying retirement can be a very powerful tool for buffering the impact of low interest rates on retirement adequacy due to the resulting higher expected asset accumulation, higher Social Security benefits, and a shortened expected period of retirement. Studies using life-cycle models can shed light on how households might change their saving, retirement, work, and investment behavior jointly in response to a low interest rate environment with the goal of maximizing lifetime welfare. Using a life-cycle model calibrated to U.S. data, the simulation in Horneff and Maurer (2018) concludes that households would accumulate less wealth, retire later, and work more in a low-return scenario, and that people with more education would alter their behavior more than people with less education. STUDIES OF HOW PERSISTENT LOW INTEREST RATES COULD AFFECT PENSION FUNDS, INSURANCE COMPANIES, ASSET MANAGEMENT FIRMS, AND SOCIAL SECURITY The potential impacts of a persistent low interest rate environment on institutions that provide retirement income and implications for retirement security are summarized in the table below.
7
POTENTIAL IMPACT OF PERSISTENT LOW INTEREST RATES ON INSTITUTIONS AND IMPLICATIONS FOR RETIREMENT
SECURITY
Assets/ Direct impacts of low interest Indirect consequences Implications for retirement security
obligations rates
($ trillion)
Public defined $8.6 Public DB plans have lowered To maintain the high discount rates In the short to medium term, the impact of low interest
benefit plans their discount rates only slightly (assumed returns) public DB plans rates through public DB plans is expected to be limited
in response to the low interest have increased allocation to equity- because the finances of public DB plans are backed by tax
rate environment, largely due like asset classes, leading to greater revenue and the benefits are usually statutorily protected.
to the unique regulatory investment risk.
environment for public pension In the long term, pension reforms will generally make
plans in the U.S. Higher contributions will be required public DB plans less generous for future employees and
if lower-then-expected returns cause make plan participants bear more risk.
Low returns will cause underfunding.
challenges in meeting existing
pension obligations. More and more public DB plans
choose to cut benefits for new
members, share risk with plan
members, and introduce DB-DC
hybrid plans.
Corporate $3.4 The reported liabilities increase Reaching for yield and risk taking is In the short to medium term, impact of low interest rates
defined because discount rates are tied discouraged by relatively strict through corporate DB plans is expected to be limited
benefit plans to market interest rates. funding requirements and the because the strict regulations on corporate DB plans
concern of accounting expense reduces the likelihood of insolvency and the benefits are
Low returns will cause volatility; the allocations to equity protected by the Pension Benefit Guaranty Corporation
challenges in meeting existing have decreased after 2006. (PBGC).
pension obligations.
Low interest rates may accelerate In the long term, the availability of secure retirement
the transition from DB to DC. income will be reduced as more corporate plans transition
from DB to DC.
Life insurance $2.6 Currently, the declining interest The incentive for reaching for yield is In the short to medium term, impact of low interest rates
companies (annuity) rates only affect the reported limited by capital requirements, through life insurers is expected to be limited because the
liabilities of new policies. A new while evidence exists for reaching for strict capital requirements for insurers reduces the
FASB guidance (effective after yield within risk-categories of bond likelihood of insolvency and the shares of annuity and life
2023) will make discount rates investments. insurance products in total retirement savings are small.
and liabilities more sensitive to
interest rate changes. Markets for annuities and life In the long term, the availability of secure retirement
insurance products with guaranteed income will be reduced if the markets for annuity and
Low returns will cause returns are likely to shrink. guaranteed-return insurance products shrink.
challenges in meeting existing
obligations for annuities and The overall profitability will be
products with guaranteed reduced because of lower interest
returns income.
Retirement $15.8 Investment returns will become Asset-management firms may seek Lower investment returns will eventually translate into
savings in (DC and IRA lower. The aggregate size of the to improve the risk-return profiles of lower retirement income that can be drawn from DC and
defined- savings) impact will depend on the their portfolios by increasing other personal retirement savings, unless employees and
contribution fund’s investment options and exposure to illiquid assets, increasing employers contribute more, or take on more risk in their
plans and the extent to which the equity the use of leverage, taking long-short investment portfolio.
asset part of the portfolio also suffers positions, exploiting "investment
management from a low-interest-rate styles". (These proposed approaches
firms environment. are not common and are subject
skepticism)
Social Security Considerably Lower interest income for The finances of Social Security are dominated by
greater than Social Security Trust fund demographic trends and tax and benefit policies.
$35.2
As the Social Security trust fund shrinks in size, the low
interest rates are not expected to have a substantial
impact on Social Security.8 Defined benefit funds and life insurance companies will face the challenge of meeting their existing obligations of guaranteed benefits with lower expected returns in the prolonged low-interest-rate environment. Public pension funds, corporate pension funds and insurance companies face different regulatory environments and respond to the low-rate environment differently. Public pension funds have the largest exposure to risky assets, partly due to their unique regulatory environment. These institutions face pressure to improve investment returns, often by investing in riskier assets. In the research literature and much of the practitioner literature, this often is referred to as a “search for yield.” The term is a slight misnomer because institutions may take on risk to seeking higher capital gains, higher dividends, or other forms of asset return that might not technically be termed “yield.” In this report we depart from the practice of several researchers and use the term “search for return” for this behavior rather than “search for yield”. According to one paper, these behaviors are more moderate among corporate pension funds and insurers for which risk-taking is discouraged by the regulatory environment. Although challenges exist, the impacts of low interest rates on retirement security through defined benefit plans and life insurers are expected by researchers and analysts to be limited in the short to medium term. The funding of public defined benefit plans is backed by tax revenue and the benefits are usually statutorily protected. Private sector pension plans in the U.S. face more stringent rules about underfunding, which discourage excessive risk- taking and reduce the risk of insolvency. Moreover, private pension benefits are protected by the Pension Benefit Guaranty Corporation (PBGC). Life insurance companies also face strict regulations that limit risks they may take and thereby limit the risk of insolvency. One implication of the research we reviewed is that if the low interest rate environment persists into the long run, defined benefit pension funds and life insurers may have no choice but to reduce the guaranteed benefits for new workers and clients. As a result, secure retirement income will become increasingly less available to retirees in the future. For defined contribution savers, reduced portfolio returns in the low-interest-rate environment will eventually translate into lower retirement income from defined contribution savings, unless they contribute more or take on more risk in their investment portfolio. Interest rates will have very limited impact on Social Security through the direct mechanism of lower returns on trust fund assets because the trust funds are expected to shrink in size and interest income will become increasingly small. Rather, the main driving forces for Social Security finances over the longer term are demographics and tax policies. To the extent lower interest rates affect economic growth, this could affect tax revenue available for Social Security, but the direct interest-rate impact will be minimal.
9
1 Motivation and introduction
After decades of declines in real and nominal interest rates, intensified by the Great Recession and COVID-19, and
monetary policy responses, the United States has been and may continue to be in a protracted period of low
interest rates. The decline of long-term rates is a global phenomenon that was largely unanticipated.
This could affect the retirement security of Americans profoundly, through impacts on investment returns, on the
capacities of governments and the private sector to finance Social Security, pension funds, and other retirement
savings mechanisms, and on the economy more broadly.
Research by academics and practitioners can shed light on the implications of the prolonged low-interest-rate
environment, but research to date has been complicated and mixed. The research has not yet been synthesized in a
cohesive, analytical, comparative manner. Our objective in this report is to provide a careful review that clarifies the
likely impacts and helps explain why different studies reach seemingly different conclusions.
Researchers, practitioners, and the media use the term “interest rate” to refer to different kinds of rates. One
important distinction is between nominal interest rates and real interest rates, which are nominal rates adjusted for
actual or expected inflation. Economists make a second distinction, between observable interest rates, which
among other things reflect transitory factors such as business cycle conditions and changes in monetary policy, and
an unobservable “neutral” interest rate that reflects longer run fundamental factors that affect investment and
savings decisions and would prevail if the economy were at full employment with stable inflation and neutral
monetary policy (neither contractionary nor expansionary). Economists also refer to this rate as the natural rate or
the long-run equilibrium rate.1 2 Much of the research we discuss on causes of the decline in interest rates and
whether it will persist examines underlying forces affecting the long-run equilibrium rate, which is the primary driver
of economic decisions that will affect retirement security. In addition, the forces driving the long-run equilibrium
rate will drive observed interest rates over the long run, even if other forces affect short run cyclical swings in
observed rates.3 In this report, when referring to interest rates we use “long run” to mean an interest rate in long
run equilibrium, and we use “long term” to refer to an interest rate on a security with longer-term duration, such as
a 30-year bond.
We cover three major topics in this literature review, in the following order.
1. Studies of whether the low interest rate environment will persist
2. Studies of how persistent low interest rates could affect the retirement security of individuals
3. Studies of how persistent low interest rates could affect institutions: pension funds, insurance companies,
asset management firms, and Social Security10 2 Studies of interest rates: What caused the decline, and will it persist? In this section we summarize and synthesize important studies that examine the driving forces underlying the decline in interest rates and discuss the possible future path of interest rates. 2.1 THE DOWNWARD TREND IN INTEREST RATES SINCE THE 1980S Both nominal and real rates have declined substantially in the past 40 years (Figure 1).4 Figure 1 NOMINAL AND REAL INTEREST RATES HAVE BEEN DECLINING FOR 40 YEARS Declining interest rates are a global phenomenon that was largely unanticipated by economic forecasters. For example, the Council of Economic Advisors (2015) showed that in the 1990s and 2000s rates declined in the U.S., Canada, the U.K., the Euro area, Korea and Japan. They also showed that the Blue Chip “consensus” forecasts of 10- year Treasury rates published in the mid-1990s through 2015 consistently were too high: forecasters did not anticipate that rates would stay so low for so long has important implications for retirement security. First, if past retirement plans and saving decisions were based on expectations that interest rates and asset returns would be higher than they are now, savers may not have accumulated enough wealth to support planned retirement spending. Second, the worldwide decline in real interest rates makes it difficult, if not impossible, to hedge interest rate risk by holding a globally diversified portfolio of interest-bearing assets. 2.2 WHAT CAUSED THE DECLINE IN INTEREST RATES? Several studies have examined the driving forces underlying the decline in interest rates in the past 40 years. These studies examine fundamental forces that drive interest rates over the long run under the assumption of neutral monetary policy. We focus primarily on the following studies, using short names (see the bibliography for full citations):
11 • Council of Economic Advisors (CEA) (2015): Long-Term Interest Rates: A Survey • Rachel and Smith (2015): Secular Drivers of the Global Real Interest Rate • IMF (2014a): Perspectives on Global Real Interest Rates We review two frameworks for understanding the commonly discussed driving forces of long run equilibrium real interest rates. The structure of the following analysis generally follows Rachel and Smith (2015) and Council of Economic Advisors (2015) (referred to as CEA (2015) hereafter). Framework 1 focuses on the connection between economic growth and interest rates, and Framework 2 looks at how shifts in saving and investment preferences affect interest rates. The two frameworks are not independent of each other but discussing them separately will help clarify the pathways through which major underlying factors affect interest rates. 2.2.1 FRAMEWORK 1: ECONOMIC GROWTH AND INTEREST RATES Slowing economic growth is a commonly cited factor behind the decline in real interest rates. How is economic growth connected with real interest rates? Productivity growth and labor force growth are two major components of long-run economic growth, both of which can affect equilibrium interest rates. Slowing productivity growth can put downward pressure on interest rates through its impact on savings behavior of households and on investment behavior of firms. Lower prospective productivity growth would cause households to lower their expected future income, because income and wages depend heavily on productivity. Facing lower expected future income, households may reduce current consumption and increase savings to boost their potential consumption in retirement, smoothing their consumption between working years and retirement years. The additional savings would increase the supply of loanable funds, creating downward pressure on interest rates. Slower expected productivity growth is also associated with fewer profitable investment opportunities for firms, which decreases the demand for investment and creates downward pressure on interest rates. Global productivity growth has been slowing over the past 30 years and the U.S is no exception. A broad measure of business sector productivity growth in the U.S. was only 1.25 percent from 2006 to 2015 and 0.5 percent from 2011 to 2015, compared to an average of over 2 percent per year from 1976 to 2005 (Fischer 2016). Economists differ on the outlook for productivity growth, with pessimistic and optimistic views. Gordon (2012, 2014) argues that long-term structural factors, including the slowing pace of educational attainment, rising inequality, and an increase in public debt will hold back future productivity growth in the U.S. Syverson (2013), Mokyr (2014), Brynjolfsson and McAfee (2014) take a more optimistic view, arguing that productivity growth will eventually pick up, albeit unevenly across the economy, largely as a result of digital technologies. As an increasing share of the population enters retirement, labor force growth will slow, putting downward pressure on U.S. economic growth. Aaronson et al. (2014) estimates that future labor growth will be reduced by about 0.25 percent by the effect of population aging. 2.2.2 FRAMEWORK 2: SHIFTS IN DESIRED DEMAND FOR INVESTMENTS AND SUPPLY OF SAVINGS The equilibrium long-run real interest rate equates desired savings and desired investment for any given prospect of long-run economic growth. Rachel and Smith (2015) shows that aggregate global savings and investment as a share of GDP have been relatively stable over the past 30 years as real interest rates declined. This implies that the decline in interest rates was caused by a combination of an increase in desired savings and a decrease in desired investment at any given interest rate.5 2.2.2.1 Desired savings Three factors affecting desired savings in the past 30 years have been widely discussed in the literature: shifting demographics, the rising level of inequality, and the global savings glut.
12 Shifting demographics. Savings rates vary through an individual’s lifetime. The path of income is generally hump- shaped (lower income in early years, higher income in prime working years, and lower income in retirement) but consumption is more stable over the life cycle: individuals of working age tend to have higher savings rates than younger individuals and older individuals. A higher share of working-age people in the population leads to a higher desired aggregate savings rate, putting downward pressure on interest rates, all else equal. The share of working- age people in the global population has risen from about 50 percent in 1970s to about 60 percent in 2010s (Rachel and Smith 2015), driven by factors such as baby-boomers entering the workforce and the fall in the proportion of young dependents.6 Rising inequality. Higher-income individuals tend to save a larger proportion of their income than lower-income individuals do. Rising income inequality will lead to a higher desired aggregate savings rate and lower interest rates, all else equal. Global savings glut. Bernanke (2005) discussed two major international forces that represented shifts in desired savings, particularly in emerging markets. First, many Asian governments have greatly increased their foreign exchange reserve as a precautionary measure after the Asian financial crisis in 1998. Second, oil producing countries have seen large increases in their savings due to the periods of high oil prices. These forces have increased the global supply of savings and placed downward pressure on global interest rates. Bernanke (2015) suggests a mixed outlook for the emerging-market savings glut, depending upon the country, while the IMF forecast (IMF 2014a) suggests the emerging-market savings glut will unwind very gradually in the near future. 2.2.2.2 Desired investment Three major trends have affected interest rates through their impact on desired investment: a decline in the price of capital goods relative to other business inputs such as labor, lower public sector investment, and a rise in risk- premia between risk-free rates and returns on capital. A fall in the relative price of capital goods. The relative price of capital goods has declined by about 30 percent in the past 40 years (Eichengreen 2015). The drop in the relative price of capital can affect total investment in two opposing ways: cheaper capital lowers the costs of investment projects reducing investment as a share of nominal GDP, while low costs encourage firms to increase real investment. Economic theory suggests that the net effect depends on the extent to which firms will substitute capital for labor in response to the declining relative price of capital. Empirical research suggests that capital investments are relatively inelastic with respect to the relative price of capital (IMF 2014a; Thwaites 2015), meaning the fall in the relative price of capital goods will lead to smaller nominal amount of investment at any given interest rate, pushing the equilibrium interest rate downward. Lower public sector investment. According to IMF (2014b), public investment as a share of global GDP has been declining since the 1980s, resulting in about a 1 percentage point decrease in the global investment-to-GDP ratio between 1980 and 2007.7 Although public investment in emerging economies picked up significantly after the Great Recession, the declining trend of public investment in advanced economies has continued. Rachel and Smith (2015) deemed the decline of public sector investment to be a secular trend but believe it has placed and will place relatively little downward pressure on real interest rates. Rise in spread between risk-free rate and the cost of capital. Our discussion has focused on risk-free interest rates in long-run equilibrium. However, business investment decisions depend upon the expected rate of return on capital, which reflects the risk of these investments and is related to the risk-free rate through a spread, or risk premium. Changes in this spread can affect the relationship between risk-free interest rates and desired investment. All else equal, if the required spread increases by 1 percentage point, then the risk-free rate will need to decrease by 1 percentage point to retain the same level of return on capital and thus maintain the same level of desired investment.8 (In the long run, all else is not equal: a lower risk-free rate would lead to lower desired savings, placing
13 upward pressure on the risk-free rate that would offset some of this impact.) Based on a weighted measure of global return on capital constructed by IMF (2014a), the global spread between risk-free rates and returns on capital increased by about 1 percent from the early 1990s through the early 2010s, placing downward pressure on equilibrium interest rates.9 Rachel and Smith (2015) showed that U.S. markets shared the same upward trend in the spread over a similar time period. 2.3 WILL THE LOW INTEREST RATE ENVIRONMENT PERSIST? The future path of the long-run equilibrium interest rate depends on how the underlying driving forces discussed in the previous section change over time. We summarize and compare studies that use quantitative methods to evaluate these major underlying factors and project the future trend of the equilibrium interest rate. We also review projections of interest rates provided by prominent economic forecasters and implied by financial markets. 2.3.1 QUANTITATIVE STUDIES OF LONG-RUN EQUILIBRIUM INTEREST RATES AND UNDERLYING FACTORS We review three prominent quantitative studies that cover major factors underlying interest rates and summarize the key results and conclusions in Table 1. The studies generally concluded that major factors causing the decline are likely to persist for the foreseeable future. Eggertsson et al. (2019) formalize and quantify the “secular stagnation” hypothesis for the low interest rate environment - a hypothesis driven by the idea that the decline may be driven by long-run persistent forces rather than by shorter-run cyclical forces - by constructing a series of overlapping generation models that decomposed the contribution of various factors. (Overlapping generation models can capture the changing behavior of consumers and other economic actors over different phases of their lives, and often are used to examine the dynamics of the macroeconomy over time.) In addition to factors we have discussed above, the models in this paper include government and consumer debt, which tend to push interest rates upward and offset the effects of other secular factors. This study focuses on domestic factors within the U.S. and does not discuss the impact of a global savings glut. Their model suggests the economy may be at a point in which real interest rates need, on average, to be negative to achieve full employment, although policy changes such as a higher inflation target, higher debt relative to GDP, or more generous Social Security benefits could avert this. In a speech at the Economic Club of New York, Federal Reserve Board vice-chair Stanley Fischer examined the relative importance of contributing factors to low interest rates using alternative-scenario simulations from the US/FRB model, one of the main econometric tools used by the Federal Reserve Board (Fischer 2016). This study estimated the effect of each factor in isolation, as shown in Table 1; because the effects can overlap, we cannot simply add the numbers together to arrive at combined impacts. He concluded that slower U.S. economic growth driven by lower productivity growth and slowing workforce growth was the largest contributor to the decline in interest rates, followed by changing demographics (in isolation), lower investment, and slowing foreign growth. He concludes that many of these factors are likely to persist, meaning that a low-interest-rate environment is likely to persist. However, he also analyzed future policy and behavioral scenarios that could lead to higher interest rates. The scenarios examined include 1) an increase in business-sector investment equal to 1 percent of GDP which caused a 0.3 percent increase in the equilibrium rate, 2) an increase in government spending of 1 percent of GDP (+0.5%), and 3) a tax cut of 1 percent GDP (+0.4%). Thus, he concluded that macroeconomic policies to increase investment or add fiscal stimulus could partly offset forces causing low interest rates to persist. Rachel and Smith (2015) take a global perspective and provide a comprehensive review of how underlying factors affect interest rates. Rather than running simulations with macroeconomic or econometric models, the authors try to quantify the impact of different factors on interest rates by drawing upon results and estimates from existing literature. Their study employs the framework of the Ramsey model to analyze the relationship between economic growth and interest rates and uses the framework of desired saving and investment to examine how shifts in saving
14 and investment preferences affect the equilibrium interest rate. (The Ramsey model is one of several frameworks economists use to study the dynamics of economic growth. It involves some simplifying assumptions but produces useful insights.) They assessed the relative importance of the underlying factors by reviewing and synthesizing existing estimates of model parameters and findings regarding determinants of investment and saving preferences. Rachel and Smith’s analysis indicates that slowing economic growth and demographic change can explain the bulk of the decline in real interest rates since the 1980s. They conclude that these underlying factors likely will persist and therefore real interest rates are likely to remain historically low in the foreseeable future, although there is great uncertainty around future trends in the underlying factors. The first study in Table 1, Eggertsson et al. (2019), estimates much bigger effects for some factors than the other studies. The most prominent difference between it and the other two studies is that it includes government debt in the model, which Eggertsson et al. (2019) estimated to have a very large positive impact on interest rates (+2.11%), requiring, as a technical matter, larger estimates of negative effects for other factors in order to explain the full decline. Still, that is not a fully satisfying explanation. It can be difficult to determine why estimates differ across studies because they use different analytic and modeling approaches. Differences in methods and results contribute to the great uncertainty in estimates of the long run equilibrium rate and how it is related to underlying factors.
15
Table 1
COMPARISON OF STUDIES ON WHETHER THE LOW INTEREST RATE ENVIRONMENT WILL PERSIST
Estimates of the importance of underlying factors
Real interest rate Slowing economic Demographic Weak investment Global saving
change to explain growth change demand glut Other factors Prospective analysis and projections
Eggertsson et 4.02% decrease Total: -2.42% Total: -3.77% Total: -0.44% Not analyzed While this trend may reverse itself, if
al. (2019) from 1970 to Total: +2.24% current projections for fertility and
2015 Slower productivity Reduction in Fall in relative price productivity hold, the natural rate of
growth: -1.90% fertility: -1.84% of investment Increase in interest will be low or negative for the
goods: -0.44% government debt: foreseeable future.
Changes in the Reduction in +2.11%;
labor share: mortality: - 1.92% A key determinant of whether interest
-0.52% Variation in rates are likely to increase is whether the
consumer debt rate of productivity growth returns to its
capacity: +0.13%. long-run rate of 2%.
Fischer (2016) No target -1.20% -0.75% -0.60% Slow growth in A variety of factors have been holding
other countries: down interest rates and may continue to
Effects of -0.3% do so for some time.
underlying factors
are estimated in But economic policy can help offset the
isolation and are forces driving down longer-run equilibrium
not additive interest rates (higher private investment,
higher gov't spending, tax cut).
Rachel and 4.5% decrease in Total: -1% -0.90% Total: -1.4% -0.25% Total: -0.45% The global neutral rate will remain low,
Smith (2015) global rates since perhaps settling at or slightly below 1% in
late 1980 Slower productivity Fall in relative price Higher inequality the medium- to long-term.
growth in of capital goods: within countries:
developed -0.5% -0.45% Underlying factors that may drag further:
economies: decline in global growth; falling relative
-0.75% Weak public price of capital; rising inequality.
investment: -0.2%
Weaker labor Underlying factors that may reverse in the
supply: -0.25% Rising risk future: demographic trend.
premium: -0.7%16
2.3.2 PROJECTIONS BY PROMINENT ECONOMIC FORECASTERS AND BY FINANCIAL MARKETS
Two kinds of forecasts suggest that the low interest rate environment will persist for an extended period – those
from prominent economic forecasters and those from financial markets. Error! Not a valid bookmark self-reference.
summarizes long-term forecasts of interest rates by government agencies, financial firms, and forecasts implied by
financial markets as reported in CEA (2015). We focus on forecasts of 10-year Treasury rates to facilitate comparison
across forecasts.
• The Federal Open Market Committee (FOMC): The FOMC provides long-run projections of the federal funds
rate by Federal Reserve Board members and Federal Reserve Bank presidents. The projected nominal 10-year
Treasury rate is constructed by adding an assumed term premium of 1 percent (CEA, 2015) to the median long-
run projection of the federal funds rate. We have updated the FOMC numbers to incorporate forecasts
prepared in early 2019.
• Congressional Budget Office (CBO): The projection is drawn from the Outlook for the Budget and the Economy
2019. We have updated the CBO numbers to incorporate forecasts prepared in early 2019.
• Blue Chip Economic Indicators: Blue Chip Economic Indicators collects forecasts for major economic variables
by more than 50 leading private-sector economists. The median 10-year projection of the 10-year Treasury rate
made in 2015 is presented in the table. We have not updated these estimates because we do not have access
to proprietary Blue Chip forecasts.
• Forward TIPS and Treasury rates: Forward rates reflect the expected future rates of the bond market
participants and other factors such as term premia to account for risk aversion. As such, they are imperfect
indicators of what markets expect about future rates but can provide insight. The table presents the 10-year
forward rates for 10-year TIPS and 10-year Treasury reported in CEA (2015) based on the market data in 2015.
To the extent that the forward rate includes a substantial term premium, market expectations of future
Treasury rates at the time of CEA’s 2015 paper would have been lower than the shown forward rate. This
buttresses the point that markets expected low rates to persist.
The column to focus on in Error! Not a valid bookmark self-reference. is the rightmost column, which gives an
estimated forecast of the real 10-year Treasury rate approximately 10 years after the date of the forecast.10
There are two ways to gain insight from Error! Not a valid bookmark self-reference.. First, we can compare the
forecasted rates to longer-run historical rates. The forecasted 10-year ahead real rates range from 1 percent to 2.4
percent, which are generally below the real rates of the 1980s and 1990s. Second, we can compare the forecasted
rates to real rates at the time of the forecasts, which were near zero. Thus, the econometric forecasts and forward
market forecasts of real rates suggest a rise over a decade of about 1 to 2.5 percent, with rates remaining below the
rates of the 1980s and 1990s (with the caveat above about term premia embedded in forward rates).17
Table 2
ECONOMIC AND MARKET FORECASTS OF 10-YEAR-AHEAD RATES
Projected 10-year Treasury rate
Projection horizon
and base year Projected Real rate assuming 2%
nominal rate inflation
FOMC-based Longer-run from 2019 3.5%* 1.5%
CBO 10-year from 2019 3.8% 1.8%
Blue Chip 10-year from 2015 4.4% 2.4%
Forward TIPS as
reported in CEA 10-year from 2015 - 1.0%
(2015)
Forward Treasury as
reported in CEA 10-year from 2015 3.1% 1.1%
(2015)
* Based on the projected federal funds rate of 2.5% and an assumed term premium of 1%.
2.3.3 UNCERTAINTY ABOUT FUTURE LONG-RUN EQUILIBRIUM INTEREST RATES
Forecasts of long-run interest rates have multiple layers of uncertainty. First, it can be difficult to estimate true real
interest rates (Laubach and Williams 2003; 2016; Hamilton et al. 2016). Second, long-run trends in underlying
factors driving interest rates and their relationships with interest rates are uncertain. Finally, financial markets may
overreact to the current environment.
2.3.4 THE IMPACT OF COVID-19 PANDEMIC
The impact of the COVID-19 pandemic on the long-run equilibrium interest rate depends on how the pandemic will
affect the underlying driving forces of interest rates discussed above. Economic theories suggest that pandemics can
put downward pressure on interest rates through reduced investment opportunities and increased preference to
save. There is emerging literature on the possible impact of COVID-19 on the long-run equilibrium interest rate.
Jordà et al. (2020), for example, examined the long-term economic consequences of previous severe pandemics and
found that pandemics can have long-lasting effects on interest rates: real interest rates decline by nearly 1.5 percent
for about 20 years and it takes an additional 20 years for the equilibrium rate to recover. The authors also point out
that the impact of COVID-19 on interest rates may be lower than the estimate based on historical events due to at
least three factors: (1) deaths caused by COVID-19 relative to the total population could be smaller due to improved
medical and public health measures, (2) the health consequence of COVID-19 is disproportionally greater for the
elderly, who are less likely to be in the labor force, and (3) massive stimulative fiscal policies in response to COVID-
19 will lead to further increases in government debt, which will place upward pressure on interest rates.
2.4 SECTION CONCLUSIONS
The decline in interest rates was not anticipated by forecasters and has been a global phenomenon with nominal
and real interest rates in almost all major economies declining over the past four decades. A consensus has been18 forming among economists that the low interest rate environment will be persistent, although interest rates may rise moderately as the effects of some short-run factors fade away. Projections of long-run equilibrium interest rates vary across studies, largely due to differences in analytic approaches and in assessment of the relative importance of underlying factors. Despite the consensus on the likely persistence of the low interest rate environment, these forecasts are subject to great uncertainty.
19
3 Studies of how persistent low-interest rates could affect the retirement security
of individuals
We summarize and compare prominent studies that examine the potential impact of the low interest rate
environment on the retirement security of individuals. To facilitate comparison across studies, we first provide an
overview of how different sources of retirement income are affected by the low interest rate environment and
review important considerations in evaluating retirement security.
3.1 SOURCES OF RETIREMENT INCOME AND HOW THEY ARE AFFECTED BY LOW INTEREST RATES
The main sources of income and resources for U.S. residents age 65 or older are Social Security, income from
earnings, pension income, income from financial assets, and real estate. The extent to which older individuals rely
on these sources varies greatly by income range. Figure 2 reproduces data from Poterba (2014), which collapses
several categories.11 It shows that Social Security is the largest income source by far for older individuals with below-
median income and remains quite important for those between the median and the 75th percentile. Income from
earnings is the most important source for individuals in the top quartile and is relatively important for the next
quartile. Income from assets constitutes about 10 percent of total income for those above median income but plays
a small role for those below the median.
Figure 2
SOCIAL SECURITY IS AN IMPORTANT INCOME SOURCE FOR THREE-QUARTERS OF OLDER AMERICANS
Percentage of Income Received From Each Source,
Individuals Age 65+ in 2013
43.7%
21.8%
Income Quartile (Average)
Highest ($78,180) 13.8%
18.1%
12.9%
Third ($26, 640) 21.1%
6.8%
56.5%
4.2%
Second ($15, 401) 6.2%
3.7%
83.5%
2.3%
Lowest ($6,756) 2.5%
2.7%
85.0%
0% 20% 40% 60% 80% 100%
Percentage of Total Income
Earnings Pension Income Asset Income Social Security
Source: Table 6 in Poterba (2014), based on March 2013 Current Population
In this section we describe the major sources of retirement income. Section 4 will examine the impact of low
interest rates on institutions that provide retirement income, through effects on the market value and cash-flow-
generating capability of the assets underlying this income.20 Social Security is the most-important source of retirement income for the majority of Americans aged 65 or older, and for about a quarter of the older population it accounts for more than 90 percent of income (Dushi et al. 2017). (Different studies examine different subgroups of the older population and use slightly different data sources, so this is not directly comparable to the figure above.) Social Security benefits are based upon lifetime earnings, calculated in a progressive manner so that benefits as a share of pre-retirement earnings (known as “replacement rates”) are greater for lower-earning individuals than for higher-earning individuals. Interest rates do not play a direct role in the calculation of benefits but could affect the ability or willingness of the federal government to pay full Social Security benefits, examined later in this report. The extent to which older people have earnings – i.e., income from work – may be surprising. However, in recent decades labor participation rates among the older population have risen, in part reflecting rising longevity (Poterba 2014).12 Payments from defined benefit pensions and annuities generally represent a secure regular source of retirement income backed by the underlying financial assets. When interest rates fall, annuity contracts become more expensive: the same amount will buy lower annuity payouts. The cost of buying the same amount of annuity income has doubled in the past 30 years because of improved longevity and the decline in interest rates. Defined benefit pension and annuity providers may change policies in response to the low interest rate environment to maintain sustainability. An important reason for this is that the current low interest rate environment was largely unexpected, as discussed in section 2. The impact of lower interest rates on institutions is discussed in section 4. Payments from defined contribution plans, such as 401(k), individual retirement arrangements (IRAs) and similar accounts depend upon the assets in those accounts. The plans are invested in financial assets such as direct investments in bonds and equities and indirect investments through vehicles such as mutual funds and ETFs. The market values of existing bonds are inversely related to interest rates. Therefore, the current low interest rate environment implies a risk of market-value loss of fixed-income assets if interest rates rise. If low interest rates persist, new bond issues will bear lower interest rates. Changes in long-run interest rates can affect stock returns through many channels. Over the long run, stock returns reflect the risk-free return plus a risk premium (the anticipated return to risk taking). With lower interest rates, stock returns will be lower unless the risk premium rises. Professional forecasters generally expect low future returns to equity, compared to historical experience.13 Factors underlying the low-interest rate environment can also affect stock returns indirectly. The anticipated slow productivity growth underlying the low interest rate environment is also a headwind for long-run stock returns: slow productivity growth may lead to lower expected and actual profit growth and in turn cause slower appreciation of market values and slower growth of dividends. Real estate may play an increasingly important role in retirement security. Housing assets are a more significant form of wealth than financial assets for most households, and mortgages are largely paid-off for individuals close to retirement age. Reverse mortgages are an important tool that can convert housing equity into cash income in retirement and can provide some hedge against declining interest rates: the proportion of home value that can be borrowed through a reverse mortgage is inversely related to the nominal interest rate. As we show below, assumptions about how real estate is used as a source of retirement income can greatly affect the estimated impact of low interest rates on retirement adequacy. 3.2 IMPORTANT CONSIDERATIONS IN EVALUATING RETIREMENT SECURITY We review three important considerations in evaluating retirement security: (1) different population subgroups, such as age cohorts and income groups, face different situations in terms of retirement planning and preparation, (2) individuals and families may have behavioral responses to shifts in the economic environment that can mitigate
21
the impact on retirement security, and (3) retirement security can be measured using different methods, each of
which has advantages and disadvantages.
3.2.1 DIFFERENCES ACROSS POPULATION SUBGROUPS
Age cohorts. Younger cohorts face greater challenges preparing for retirement than older cohorts. First, younger
generations have longer expected lifespans. In the low interest rate environment, the younger generation will need
to save for a longer expected retirement period with a lower expected return. Second, defined benefit retirement
plans are much less common for the younger cohorts than for Baby Boomers. They will need to rely more upon their
own investments and associated investment risk. If they want the secure annual income that defined benefit plans
provide, they can purchase annuities, but annuity plans are more expensive to purchase for young people now than
they were for Baby Boomers due to increased longevity and lower expected returns - it is nearly twice as expensive
to buy a given annuity now as it was 30 years ago.14 Furthermore, annuity income is more sensitive to changes in
interest rates when expected lifespans are longer than when they were shorter. On the other hand, the challenges
described above may be mitigated by the lower borrowing costs younger cohorts face in a low interest rate
environment.15 Income groups. The impact of the persistent low interest rate environment differs across income
groups. Low-income households are generally affected by persistent low interest rates to a lesser extent than high-
income households because a larger share of their retirement income is from Social Security and they generally
have lower financial assets. Also, the longer life expectancy of high earners makes their retirement income more
sensitive to changes in expected asset returns (Olshansky et al. 2012).
It generally is easier for high-income individuals to mitigate the impact of low interest rates through behavioral
changes without making major lifestyle changes: because consumption expenditures of higher-income individuals is
a smaller share of income than it is for lower-income individuals, particularly for necessities: it may be easier to add
to savings by scaling back consumption of non-necessities rather than by cutting back on necessities or working
longer.16
Higher income households are also more likely to own more housing that can be converted into retirement income.
3.2.2 BEHAVIORAL FACTORS
Individuals and families may choose to improve their retirement security through changes in savings behavior,
housing consumption, and retirement timing in response to shifts in the economic environment, particularly interest
rates. These behavioral responses vary across population groups: Low income groups are less likely to own real
estate, have less room to increase their savings rate, and may have less flexibility over when to retire.
3.2.3 MEASURING RETIREMENT SECURITY17
Three common methods of measuring retirement security are replacement ratios, success or failure rates, and
wealth ratios. Each has advantages and disadvantages for measuring retirement security.
Replacement ratios calculate post-retirement income as a percentage of a measure of pre-retirement income and
compare this to a target percentage that is deemed to be “adequate.”
• Advantages: The primary advantages of replacement ratios for analysis of retirement income adequacy are
that they are easily explained, can be compared over time, and may be used for individuals, groups of
individuals or the entire retiree population.
• Disadvantages: However, there is no universally agreed-upon definition or consensus on what constitutes
an adequate replacement ratio or on how to adjust for differences in individual circumstances.You can also read
